High-Level Sound Classification in the ESOUNDMAPS Project

摘要: ESOUNDMAPS is an ongoing research program that aims in developing a wireless audio sensor network (WASN) and deploying it at the surrounding environmental area of Technological Institute Piraeus Attica, Greece. The proposed WASN will be used for monitoring to future assessing impact human generated noise wildlife. Collected sound samples forwarded by nodes central server where they automatically evaluated with respect their identity; maps based on evaluation these samples. High-level classification defined herein as act classifying sample three broad categories namely anthropogenic, biophysical (other than human) geophysical sounds. In this paper we present integrated platform includes denoising using wavelets, feature extraction from Gaussian mixture modeling features, powerful two-layer neural classifier automated high-level incoming Classification results, obtained digital samples, exhibit outstanding accuracy (sometimes reaching or exceeding 98% correct vs. incorrect estimates), thus demonstrating feasibility approach realistic environments.